Abstract:Traditional artificial potential field methods face limitations in multi-UAV cooperative path planning, such as swarm instability and local minima. To address these issues, this study proposes a cooperative obstacle avoidance method that integrates bidirectional coupling of formation potential fields with dynamic omnidirectional disturbance potential field design. To mitigate swarm motion instability, this study establishes a formation-level attractive-repulsive cooperative mechanism through a coupling strategy of dynamic reference point generation and inter-UAV state perception, effectively reducing the probability of trajectory conflicts. To resolve issues such as unreachable targets and local oscillations, it designs a dynamic omnidirectional disturbance potential field, incorporating a collinear real-time detector, dynamic reference axis generator, and adaptive disturbance regulator, enabling escape from local equilibrium points while maintaining smooth trajectories. Monte Carlo experimental results demonstrate that the proposed method achieves a target arrival rate of 91% in scenarios with dense, mixed dynamic obstacles, exhibiting higher reliability compared to benchmark algorithms.